Complex network modeling is an elegant yet powerful tool to delineate complex systems. Hierarchical clustering of complex networks can readily facilitate our comprehension of the higher order organizations of complex systems. Among all the complex network models, bipartite network is an essential part. In this paper we present a multiobjective optimization based hierarchical clustering algorithm for bipartite networks. In doing so, we first devise a similarity index whereby a bipartite network is mapped into a monopartite network. We further put forward a multiobjective optimization model for monopartite network clustering. Finally we develop an agglomerative method for deriving the hierarchical tree structure of the original bipartite network. To evaluate the effectiveness of our proposed bottom-up hierarchical clustering algorithm, we carry out experiments on ten bipartite ecological networks. We also compare our algorithm with one state-of-the-art bipartite network clustering algorithm and one highly efficient hierarchical network clustering method. Experimental comparisons show the efficiency of our proposed algorithm for hierarchical clustering of bipartite networks. By further analyzing the hierarchical trees derived by our proposed algorithm we find that our obtained trees are biologically appealing and could have potential implications for species classification.

The air traffic control paradigm is shifting from sector-based operations to flow-centric approaches to overcome sectors’ geographical limits. Modeling and predicting intersecting air traffic flows can assist controllers in flow coordination under the flow-centric paradigm. This paper proposes a flow-centric framework – TEMPT: Text-Enriched air traffic flow Modeling and Prediction using Transformers – to identify, represent, and predict intersecting flows in the airspace. Firstly, nominal flow intersections (NFI) are identified through hierarchical clustering of flight trajectory intersections. A flow pattern consistency-based graph analytics approach is proposed to determine the number of NFIs. Secondly, in contrast to the traditional traffic flow feature representation, i.e., numerical time series of flights, this paper proposes a text-enriched flow feature representation to intuitively describe the “flow of flights” in the airspace. More specifically, air traffic flow features are described by a “text paragraph” composed of the time and flight sequences transiting through the NFIs. Finally, a transformer neural network model is adopted to learn the text-enriched flow features and predict the future traffic demand at the NFIs during future time windows. An experimental study was carried out in French airspace to validate the efficacy of TEMPT using one-month ADS-B data in December 2019. Prediction results show that TEMPT outperforms the competitive air traffic flow modeling and prediction approaches: time-series-based Transformers, Long Short-term Memory (LSTM), and Graph Convolutional Networks (GCN), as well as aerodynamic trajectory simulation-based prediction and the historical average.

Air traffic is operated in an air traffic network (ATN) environment. It is pertinent to improve the robustness of ATNs as they are frequently exposed to manifold uncertainties which can break down their functioning components. Existing studies improve the robustness of an ATN by either rewiring its links or adding more ones. In this paper we discover the robustness paradox phenomenon in ATNs. Specifically, we claim to improve the robustness of an ATN by removing its links. In order to determine the links whose removal can improve an ATN's robustness, we develop a bi-objective optimization model with one objective maximizing the network's robustness and the other one minimizing the number of links to be removed. We further apply and modify a non-dominated sorting genetic algorithm (NSGA-II) to optimize the developed model. We then carry out experiments on nine real-world ATNs to validate the effectiveness of the proposed idea. We also compare the modified NSGA-II algorithm against NSGA-III, and MODPSO, which are famous and efficient multiobjective evolutionary algorithms. Experiments indicate that NSGA-II outperforms the compared algorithms and that robustness paradox phenomenon does exist in ATNs. This work provides a new perspective for aviation decision makers to better design and manage ATNs.

A complex network in reality often consists of profuse components which might suffer from unpredictable perturbations. Because the components of a network could be interdependent, therefore the failures of a few components may trigger catastrophes to the entire network. It is thus pivotal to exploit the robustness of complex networks. Existing studies on network robustness mainly deal with interdependent or multilayer networks, little work is done to investigate the robustness of multipartite networks which are an indispensable part of complex networks. Here we plumb the robustness of directed multipartite networks. To be specific, we exploit the robustness of bi-directed and unidirectional multipartite networks in face of random node failures. We respectively establish cascading and non-cascading models based on the largest connected component concept for depicting the dynamical processes on bi-directed and unidirectional multipartite networks subject to perturbations. Based on our developed models, we respectively derive the corresponding percolation theories for mathematically computing the robustness of directed multipartite networks subject to random node failures. We unravel the first-order and second-order phase transition phenomena on the robustness of directed multipartite networks. The correctness of our developed theories have been verified through experiments on computer-generated as well as real-world multipartite networks.